Alibaba cut the cost of feeding an hour of audio to AI by 98%.
Alibaba's Qwen team released Qwen3.8-Omni-Flash on Friday, a model that reads text, images, audio and video. Audio input now costs over 98% less per hour than the previous Qwen omni model. It holds a million tokens of context, but Alibaba did not publish the weights.
Why it mattersCheap audio is what turns voice from a demo into something running all day. Price, not quality, now decides which model gets to listen to you.
▶ Related video: Qwen3.8 Omni Flash: Fed It Images, Video, and Audio All at Once
Read in the app — free, in 9 languages
Related stories
Google's new voice model switches between 97 languages on its own.2026-09-19Salesforce trained its own reasoning model so your data stays put.
2026-09-15Shanghai's top open agent model arrived with no announcement at all.
2026-09-15DeepSeek is testing its first voice mode on a slice of users.
2026-09-14Xiaomi's open model picks one voice out of three talking at once.
2026-09-14